‘Lights Out’ Data Centers Are Now Possible, But Hard To Implement
With the rapid innovation of machine learning and artificial intelligence tools, some companies are now trying out the idea of fully automating data centers for periods of time. The idea of a “lights out” fully autonomous data center operation has been spoken about for years, but only now – thanks to AI and ML innovation – can this become a reality.
For example, data center company said its business premise was based on providing “lights out” data centers. Although EdgeConneX isn’t entirely “lights out” yet, the company has remote control physical security and uses remote monitoring and sensors to track operations, meaning employees are deployed to sites only when needed.
The goal is to get as many energy-efficient and technology effective solutions inside a data center that can be remotely monitored by data center infrastructure management (DCIM) software. By using AI, ML and potentially robots to eliminate the need for employees in the physical space, data centers could optimize oxygen levels and physical security to reduce energy, completely cut out the need for lighting, create more efficient cooling designs, and increase IT rack capacity and heights.
However, the ability to various AI and ML tools to seamlessly interoperate with each other algorithmically for critical decision making is a major challenge, not to mention the social and economic impacts.